U.S. patent number 8,150,621 [Application Number 12/386,181] was granted by the patent office on 2012-04-03 for command and control of autonomous surface vehicle.
This patent grant is currently assigned to The United States of America as represeneted by the Secretary of the Navy. Invention is credited to David B. Hanger, Craig A. Phillips.
United States Patent |
8,150,621 |
Phillips , et al. |
April 3, 2012 |
Command and control of autonomous surface vehicle
Abstract
An operating system is provided for controlling an unmanned
vehicle. The system includes a stratified plurality of instruction
layers, a behavior axiom block and a set of operation parameters.
The instruction layers are substantially arranged in descending
priority order. Each layer provides an information signal to either
an adjacent descending layer or an operation device on board the
unmanned vehicle. The behavior axiom block provides an independent
protocol signal to a first instruction layer in said stratified
plurality. The operation parameters provide an environmental
condition that neighbors the unmanned vehicle to a second
instruction layer. Preferably, the behavior axiom block includes
prioritization adjustment to an instruction layer for overriding
the information signal from an adjacently ascendant layer, such as
by an interrupt signal.
Inventors: |
Phillips; Craig A. (King
George, VA), Hanger; David B. (King George, VA) |
Assignee: |
The United States of America as
represeneted by the Secretary of the Navy (Washington,
DC)
|
Family
ID: |
45877430 |
Appl.
No.: |
12/386,181 |
Filed: |
April 7, 2009 |
Current U.S.
Class: |
701/411;
701/448 |
Current CPC
Class: |
G05D
1/0206 (20130101); G05D 1/0088 (20130101) |
Current International
Class: |
G01C
21/00 (20060101) |
Field of
Search: |
;701/1-2,209-211 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
Scokaert, P.O.M. et al., "Suboptimal Model Predictive Control",
IEEE Trans. on Automatic Control, Mar. 1999, v. 44, No. 3, pp.
648-654. cited by other .
Ohlmeyer, E. J. et al., "Generalized Vector Explicit Guidance", J.
Guidance and Control, Nov.-Dec. 2006, v. 29, No. 2, pp. 261-268.
cited by other.
|
Primary Examiner: Elchanti; Hussein A.
Attorney, Agent or Firm: Thielman, Esq.; Gerhard W.
Government Interests
STATEMENT OF GOVERNMENT INTEREST
The invention described was made in the performance of official
duties by one or more employees of the Department of the Navy, and
thus, the invention herein may be manufactured, used or licensed by
or for the Government of the United States of America for
governmental purposes without the payment of any royalties thereon
or therefor.
Claims
What is claimed is:
1. An unmanned vehicle operating a control process, said process
comprising: a stratified plurality of instruction layers
substantially arranged in descending priority order, said plurality
further including: a commander intent interpretation layer issuing
a strategic constraint, a mission design layer issuing a boundary
horizon; an environmental assessment layer providing
instrumentation information, a route planning layer that issues a
navigation parameter, a navigation layer that issues a heading
direction for the unmanned vehicle, and a control layer that issues
at least one of a steering command and a throttle command for the
unmanned vehicle; a behavior axiom block for providing an
independent protocol signal to an instruction layer in said
stratified plurality; and a set of operation parameters for
providing an environmental condition neighboring the unmanned
vehicle to said environment assessment layer as said
instrumentation information.
2. The vehicle according to claim 1, wherein: said interpretation
layer provides said strategic constraint to said mission design
layer and said environmental assessment layer, said mission design
layer provides said boundary horizon to said route planning layer,
said environmental assessment layer provides said instrumentation
information to said route planning layer, said route planning layer
provides said navigation parameter to said navigation layer, said
navigation layer provides said heading direction to said control
layer, and said control layer provides said steering command to a
steer controller and said throttle command to an engine
controller.
3. The vehicle according to claim 1, wherein said behavior axiom
block includes prioritization adjustment to said instruction layer
for overriding said information signal from an adjacently ascendant
instruction layer.
4. The vehicle according to claim 1, wherein said boundary horizon
from said mission design layer includes: a path along which the
unmanned vehicle travels; a waypoint at a forward edge of said
boundary horizon; and an instant directional heading corresponding
to said waypoint.
5. The vehicle according to claim 1, wherein said order corresponds
to substantially decreasing time interval for updating said
information signal from said each layer and said adjacent
descending layer.
6. The vehicle according to claim 1, wherein said behavior axiom
block includes prioritization adjustment to a third instruction
layer for overriding said information signal from a fourth
instruction layer adjacently ascendant to said third instruction
layer.
7. The vehicle according to claim 1, wherein said behavior axiom
block includes a safety protocol directive.
8. The vehicle according to claim 1, wherein said protocol signal
includes an interrupt signal.
9. The vehicle according to claim 1, wherein at least one of said
instruction layers includes an algorithm for collision
avoidance.
10. The system according to claim 9, wherein said algorithm
includes an instruction to adjust a heading of the unmanned vehicle
to circumvent said obstruction location.
11. The system according to claim 9, wherein said algorithm
maintains an established distance from said obstacle.
12. The system according to claim 1, wherein said environmental
condition further comprises at least one of local time, inertial
reference, global positioning system signal, speed, acceleration,
relative bearing, goal position, and obstruction location.
Description
BACKGROUND
The invention relates generally to flexible command and control of
an autonomous surface vehicle. In particular, the invention
provides a stratified structure of instructions to achieve a
mission objective operating within constraint protocols.
Conventional operational methods employ remote control signal
devices provided by a human that views sensor information from the
unmanned vehicle or from other sources to send steering commands to
the vehicle. Some limited autonomy is available for situations
without obstacles, traffic, or enemy in which waypoints are issued
to the surface vehicle follow with a simple autopilot on board. The
vehicle uses Global Positioning System or a similar system to hold
the boat on bearing to the next waypoint. Commercial boat
autopilots are available for this purpose for commercial and
recreational boating applications to reduce human workload. For
these systems, human monitoring remains necessary in the event of
traffic or obstacles. For such situations or when the weather
obscures visibility or interferes with stability, direct human
control of the vehicle steering is required to ensure operational
safety and achievement of the vehicle's mission.
An example of where direct human intervention is required is the
case of steering relative to an oncoming wave to prevent rollover.
Operations such as docking or rendezvous with a command platform
all require direct human control of the ship steering. In a
situation where the USV supports combat operations, there can be
traffic present (including both friendly and hostile) which require
human intervention to direct the activities of the vehicle by
remote control. In addition, complex missions such as intercepting
a potentially hostile incoming boat would require direct human
control via a remote link.
Some autonomy is available in missile and aircraft autopilot and
missile guidance systems. Aircraft and missile autopilots deal with
narrowly defined missions such as stabilization of the aircraft or
execution of a commanded turn to a new heading. These automated
functions are fairly limited in nature and are designed to work in
a rather scripted process.
The greatest amount of autonomy in aircraft systems occurs in the
microburst recovery systems for commercial aircraft. Because the
limited time required to respond stresses the human reaction time,
consensus is developing of the utility to provide limited autonomy
to the system to fly the vehicle out of the microburst. This
represents a very scripted and optimized flight procedure.
Trajectory guidance for an autonomous land attack cruise missile
follows a scripted mission without significant flexibility. This
limits autonomous operation to a fire-and-forget weapon such as the
Tomahawk cruise missile, rather than an unscripted reconnaissance
platform such as the Global Hawk aircraft that requires real time
flexibility.
Current methods for controlling an unmanned surface vehicle require
increased manning requirements for the command platform operating
that surface vehicle. Additionally, the current approach to remote
control operation of unmanned vessels exhibits decreased
functionality during certain periods because the human operators
degrade by fatigue or lack of trained personnel. There are also
limitations on the mission because of the limited human
capabilities. Advanced automatic systems are anticipated be able to
pilot the ship in inclement weather conditions better than human
operated systems. This arises from the ability to design the system
to use sensor input rather than organic feedback to a human
operator such as Doppler measurement of water speed and from the
faster response time of automated systems.
SUMMARY
Conventional unmanned vehicle control systems yield disadvantages
addressed by various exemplary embodiments of the present
invention. In particular, various exemplary embodiments provide an
architecture for the Command and Control (C.sup.2) of an autonomous
unmanned ship or surface vehicle with a minimum of human
intervention. The Stratified Horizon Control herein provides an
architecture for creating an algorithm that interprets the highest
level of commander's orders in a linguistic format as might be
given to a human operating an equivalent vehicle and autonomously
interprets the commanders orders and develops the various levels of
controls to ultimately steer and control the speed of the surface
vehicle to achieve its commanded mission. This algorithm provides
safe, reliable, and effective execution of the commander's intent
without increased manning requirements.
Various exemplary embodiments provide an operating system for
controlling an unmanned vehicle. The system includes a stratified
plurality of instruction layers, a behavior axiom block and a set
of operation parameters. The instruction layers are substantially
arranged in descending priority order. Each layer provides an
information signal to either an adjacent descending layer or an
operation device on board the unmanned vehicle.
The behavior axiom block provides an independent protocol signal to
a first instruction layer in said stratified plurality. The
operation parameters provide an environmental condition that
neighbors the unmanned vehicle to a second instruction layer. In
various exemplary embodiments, the behavior axiom block includes
prioritization adjustment to an instruction layer for overriding
the information signal from an adjacently ascendant layer, such as
by an interrupt signal.
BRIEF DESCRIPTION OF THE DRAWINGS
These and various other features and aspects of various exemplary
embodiments will be readily understood with reference to the
following detailed description taken in conjunction with the
accompanying drawings, in which like or similar numbers are used
throughout, and in which:
FIG. 1 is a block diagram view of a strategic horizon control
system;
FIG. 2 is a plan view of an operation under a mission design
layer;
FIG. 3 is a block diagram view of operational parameter
reception;
FIG. 4 is a block diagram view of operational parameter
integration;
FIG. 5 is a plan view of a navigation grid pattern;
FIG. 6 is a plan view of a horizon network in the FIG. 2
operation;
FIG. 7 is a first plan view of an operational mission route;
FIG. 8 is a second view of the operational mission route;
FIG. 9 is a plan view of a first navigation guidance scenario;
FIG. 10 is a plan view of a second navigation guidance
scenario;
FIG. 11 is a plan view of a third navigation guidance scenario;
FIG. 12 is a plan view of a fourth navigation guidance
scenario;
FIG. 13 is a block diagram view of a route planning circuit;
FIG. 14 is a plan view of a composite navigation scenario;
FIG. 15 is a graphical view of optimizer selected time;
FIG. 16 is a tabular view of a complex object map;
FIG. 17 is a plan view of an object avoidance mission;
FIG. 18 is a tabular view of a selection matrix for obstacle
avoidance;
FIG. 19 is a graphical view of down and cross-range distances for
collision avoidance during a docking mission;
FIG. 20 is a block diagram view of a PID controller; and
FIG. 21 is a graphical view of a yaw rate response.
DETAILED DESCRIPTION
In the following detailed description of exemplary embodiments of
the invention, reference is made to the accompanying drawings that
form a part hereof, and in which is shown by way of illustration
specific exemplary embodiments in which the invention may be
practiced. These embodiments are described in sufficient detail to
enable those skilled in the art to practice the invention. Other
embodiments may be utilized, and logical, mechanical, and other
changes may be made without departing from the spirit or scope of
the present invention. The following detailed description is,
therefore, not to be taken in a limiting sense, and the scope of
the present invention is defined only by the appended claims.
FIG. 1 shows an exploded view of a plan for Stratified Horizon
Control (SHC), which represents an important aspect in various
exemplary embodiments in controlling an unmanned vehicle. Behavior
axioms are contained within a behavior axiom block (BAB) 110,
operating parameters 120, preemptory prioritization adjustor 125 as
adrenaline parameter with interrupt control. The BAB 110 includes
safety protocols. The operating procedures 120 include group
operating pictures (GOP) and individual operating pictures
(IOP).
The plan 100 includes a series of levels for Stratified Horizon
Control (SHC) 130. The levels from strategic and long-term for
decision to specific and immediate begin with commander's intent
interpretation layer 140 that yields global quantitative goals and
constraints 145. Next, mission design 150 that yields horizons for
lower controls, path constraints and horizon goals 155, and
environmental assessment 160 at the same level of detail as the
mission design 150 but subject to information received from the
operating procedures 120. The assessment 160 yields navigation
constraints 165.
Next, route planning layer 170 that yields optimized navigation
parameters 175. Next, navigation layer 180 that yields heading rate
or heading schedule commands 185 to achieve goal objectives and
avoid collisions. Finally, control layer 190 that provides steering
and throttle commands 195. Time constraints for integrated
decisions and command implementation range from interpretation 140
involving hundreds of minutes or several hours, mission design 150
involving tens of minutes, route planning 170 involving minutes,
and navigation 180 involving tens of seconds, down to control
commands 190 that involve milliseconds to seconds.
FIG. 2 shows a plan view of a mission implementation using the SHC
130 as an exemplary mission planning layer 200 that includes
constraints from the mission design 150 and environment assessment
160. A mission goal 210 can be identified by the human overseer and
is disposed beyond a land mass 220 demarcated by a coast littoral
225. Waypoints 230 together with their directional bearing arrows
denote paths marked at horizons 240 that slide along the route
planning layer 170. Goals 155 from the mission design layer 150
include the horizons 240. Obstacles denoted by individual points
within the dash oval 250 along with a larger, more complex obstacle
260 represent entities to be avoided while navigating towards the
goal 210. The navigation constraints 165 include a global survey of
objects and hazards from on-board and off-board sensors. The
navigation parameters 175 include optimized guidance parameters.
The heading commands 185 include acceleration commands to achieve
goal and avoid collisions that control the steering and throttle
commands 195.
The unmanned surface vehicles must operate relative to a parent
command vessel and with other peer surface vehicles. FIG. 3 shows a
block diagram 300 for implementing the SHC 130 relative to
multi-vehicle operation within the operating procedures 120 for an
unmanned surface vehicle (USV). Individual USV C.sup.2 architecture
310 includes vehicle management, and sensors that communicate with
the IOP, which interface with the SHC 130. GOP 320 communicate with
the parent or mother combat system (MCS) 330 that hosts the launch
platform, and also associates with the SHC 130. Off Board Assets
340, Sensor Information 350 and Sensing Requests 360 supply
information or instructions to the MCS 330. Optimal Sensor
Allocation 370 provides supplemental information to the
architecture 310.
FIG. 4 shows an expanded block diagram 400 for the operation of an
individual vehicle using SHC 130. The USV platform 410 employs the
C.sup.2 architecture 310, including the IOP 420. Together with the
GOP 320, the IOP 420 provides information to the SHC 130 for
guidance to the platform 410. A sensor suite 430 provides
environmental and operational information to the platform 410, such
as radar, infrared camera, inertial measurement unit (IMU), global
positioning system (GPS), etc. A vehicle management system 440
provides status on power, system health and communication.
Commands 450, provided by the human operator and in conjunction
with the SHC 130, operate to provide group controls 455 to the
vehicle management 440. Other communications include sensing
requests 460 from the IOP 420 to the suite 430, returning with
sensor measurements 465, such as presence of local entities with
tracks and classification, wave measurements, wind speed, etc.
Information 470 exchanges between the operating procedures 120 and
the vehicle management 440. Information 475 also exchanges between
the operating procedures 120 to the SHC 130. Sensing requests 480
can issue from the vehicle management 440 to the suite 430. The
vehicle management 440 provides status updates 485 to the SHC 130,
which can pass requests 490 to the suite 430 and throttle/steering
commands 495 to the platform 410.
The Stratified Horizon Control (SHC) 130 represents an algorithm
that divides the control processes from interpretation of the
commander's intention through the actual commands to the steering
and speed control into a set of instruction strata. The highest
stratum involves the broadest decision making and interpretation
and is called the commander's intent interpretation layer 140. This
layer deals with the more strategic aspects of the
command-and-control mission
The interpretation of the commander's intent involves the
definition of the mission, constraints, and acceptable levels of
risk. The decision making at this level has the longest (time)
horizon for consideration. Having to deal with entire missions,
decision-making may extend to the hundreds of minutes. The lowest
level of the SHC 130 provides the control outputs to the actuators
on the USV platform 410. Typically these are concerned with the
desired rudder angle and the desired throttle control though
alternate actuator arrangements may be accommodated in this
approach. This lowest level of control deals with such inputs as
heading or heading angle rate to produce the desired actuator
command. The time horizon on this level is very short and deals
with decisions affecting hundreds of milliseconds into the future.
This architecture allows the implementation of existing
off-the-shelf components such as the use of existing autopilots for
the lowest level.
Between this uppermost layer and the lowest, the stratification of
the control may assume many forms. Descending through these layers
of the exemplary architecture illustrated in FIG. 1, the planning
becomes more detailed and quantitative and with shorter time
horizons. The entire system is designed for adaptability, and thus
the horizons for each layer can be determined by the SHC 130
process. The level for setting the horizons can typically be
performed in the mission design layer 150.
The functions of each of these layers are discussed in detail in
the following sections. These layers have set horizons that are
updated at periodically at an interval much less than their above
horizons (typically one-tenth of the immediately above horizon). At
each update the solution can be completely redesigned with the
current states as the initial conditions. Thus, the mission
designed at the start of a mission may change considerably by the
time of mission completion. Given sufficient environmental
information and no surprises, the SHC 130 provides a stable mission
definition but the redefinition can be established for adaptability
upon receipt of new information.
Another important aspect of the SHC 130 is the Behavior Axioms
Block (BAB) 110, which exists outside of the stratified layer and
is updated at a very high rate, typically on the order of the
control layer 190. That horizon can be relatively brief and
confront issues where any of the BAB 110 may be violated. The
behavior axioms deal with the survivability of the vessel and the
safety of friendly and non-combatant ships and individuals. The BAB
110 conducts its own calculations over its horizon to determine
whether any behavior axioms may be violated.
The BAB 110 has the ability to interrupt any layer currently in
process by issuances of interrupt message. This interrupt message
instills an adrenaline factor in each control factor which can
affect its actions. This factor orients the control layer 190 away
from mission success to satisfying the BAB 110, which also dictates
when the axioms are not in danger of violation and allow resumption
of normal operations for the SHC 130. At this point, the
commander's intent interpretation layer 140 reinitializes the SHC
130 again from the current vehicle states and environmental
conditions.
The commander's intent interpretation level 140 is designed to take
linguistic based commands and translates these into mission design
parameters. The interpretation of the Commander's Intent involves
the definition of the mission, constraints, and acceptable levels
of risk. The decision making at this level has the longest (time)
horizon for consideration. Having to analyze entire missions, the
interpretation level 140 may produce decisions out to the hundreds
of minutes. The outputs of this level include the definition of the
type of mission such as a patrol of specified region in use of
particular sensor suite. Once the type of mission is specified then
the mission parameters can be defined.
For example, for a patrol using a specific sensor suite, this level
would provide the boundaries of the region, spacing and interlacing
of vehicle sweeps, completion time for the patrol, decision node
and reporting points, and the acceptable mission risk of the
mission. This last parameter deals with the level of urgency
associated with the mission and is important for regulating the
adrenaline factor used in the lower levels. This factor is used to
trade-off risk for probability of mission success in the mission
design parameters and for the threshold for when the BAB 110
overrules the planned mission design 150. An example of such a
trade-off might be a very high priority mission that follows a path
that hits waves at an approach angle that would be avoided in a
routine mission. This level depends on the use of embedded
knowledge of naval operations to interpret the linguistic commands
from the commander. This level employs expert systems to allow
access to the level of knowledge that an expert user of the
particular boat and experienced sailor might possess of boat and
navy operations.
The waypoints 230 between which the platform 410 operates can be
described by nodal geometry. FIG. 5 shows a plan view diagram of a
navigation grid 500. Nodes 510 are defined by grid position and
approach direction and can be arranged in a rectangular pattern
separated by a cell width 520 having length d in orthogonal
directions. The platform 410 can be directed to follow a directed
cycle 530 forming an octagonal ring of inward arcs. Outgoing arcs
540 correspond to approach direction .+-.45.degree. directed away
from the cycle 530.
FIG. 6 shows expanded plan view 600 from FIG. 2 of mission planning
layer 150. A sliding window "world view" 610 provides an
observation region from which a moving platform 410 can operate
from waypoints 230 along concatenated horizons 240 towards the
objective 210. Expansion 620 of an exemplary horizon 240 shows a
detail horizon 630 with the corresponding waypoint 230 disposed at
a horizon boundary 640. To detour obstacles 650 and escape a
hostile vessel 660, the platform 410 travels along a navigation
line 670 towards a horizon goal at bearing 680.
The mission design layer 150 receives the output of the commander's
intent interpretation layer 140 and converts these into quantified
mission parameters. The outputs of this layer may typically
include: contemporary (and estimates of future values) horizons
(time values) for each of the layers and the desired position and
heading 230 of the vessel platform 410 at the next horizon 240.
Planning in this layer occurs at length and time scales where the
dynamics of the vessel are negligible. For the analysis at this
level, the platform 410, launch platform 330 may be considered to
turn instantaneously and paths consist of straight line segments
between nodes 510. In this layer the nodes 510 are used to specify
the path 530 or 540.
The selection of nodes 510 includes the avoidance of known large
obstacles 650 such as coasts or reefs. Smaller obstacles such as an
isolated ship or rock are not considered at this level. The
discrimination between which obstacles are included at this level
of analysis is made based on maximum dimension of the object. For
this determination, groups of closely spaced objects 250 can be
combined into a single larger object through clustering algorithms
1700. Consequently, the composite larger object or regions are
avoided by mission design 150.
The spacing of nodes that demarcate setpoints on the horizons that
are applied to the lower levels and are determined by the type of
mission interpreted from the commander's intent interpretation
level 140. As an example, if the commanding officer desires to
leave port, transit the Atlantic, and then dock at another port,
the nodes can be spaced close together in the port regions and
relatively far apart in the open ocean areas. Correspondingly, the
horizons 240 applied to the deeper layers would be smaller in the
ports and larger in the open ocean areas. FIG. 2 for the mission
design layer 150 shows the development of horizons for the route
planning layer 170 and the selection of node points by the mission
design layer 150. The extent of horizon of the route planning layer
170 represents the horizon for the mission design layer 150 which
the latter self-selects based on the goal.
The mission design layer 150 develops the horizons for the lower
levels and develops the "local world view" for the GOP 320 and IOP
420. The interrelationship of these layers is shown in FIG. 6 for
interactions of the SHC 130. The purpose of the route planning
layer (RPL) 170 is to compute the nominal parameters used in a
navigation algorithm, as in the navigation layer 180. The selection
of these parameters is based on optimizing a selected performance
function over the horizon of the RPL 170. This timeframe elapses
typically on the order of tens of seconds. The specific horizon 630
of the RPL 170 in operation is selected by the mission design layer
150. The RPL 170 represents a simulation based algorithm that uses
a simulation of the vehicle to determine the navigation algorithm
parameters. The navigation parameters 175 are based on minimizing
the time to the desired position and heading at the current RPL
horizon. The desired position and heading on the RPL horizon are
provided by the mission design layer 150.
FIG. 7 shows a plan view of a navigation grid 700 for an example
reconnaissance mission as an exemplary embodiment for the mission
planning design layer 150. An initial waypoint 710 provides a start
position for the platform 410, which is assigned to reconnoiter a
patrol area 720 containing obstacles 650, while avoiding keep-out
zones 730. The platform 410 begins from the initial waypoint 710
along an ingress path 740 to the patrol area 720. Upon arrival, the
platform 410 proceeds to maintain station in a loiter cycle 750
within the patrol area 720 to survey an object 760 under
observation. Upon completion of this task, the platform 410 returns
to its initial waypoint 710 by an egress path 770 while again
avoiding obstacles 650 and the keep-out zones 730.
FIG. 8 shows an expanded plan view of the navigation grid 800. The
ingress and egress paths 740, 770 include demarcation horizons 810
that mark nodes 510 and course change positions. At the entry of
the patrol area 720, the path 740, 770 includes goal horizons 820
to indicate time-and-position goals.
FIG. 9 shows a plan view 900 for a first guidance level competency
for the navigation layer 180 with guidance to commanded time,
position and heading at the next horizon 240. The platform 410 is
disposed at an initial horizon boundary 910 corresponding to
interval i and travels along a path 920 towards an adjacent horizon
930 at the next interval i+1, denoted by cross-hair oval, while
avoiding obstacles 650. An exemplary north-east-down-local (NEDL)
frame 940 references global compass directions oriented on
first-quadrant Cartesian directions north 950 and east 960.
Alternative frames using different directions and/or for other
quadrants can be employed.
FIG. 10 shows a plan view 1000 for a second guidance level
competency for the navigation layer 180 guidance to intercept a
target with offset rendezvous. The platform 410 is disposed at an
initial horizon boundary 1010 corresponding to interval i towards a
target vessel 1020 to be intercepted. An adjacent horizon 1030 at
the next interval i+1 corresponds to a position not directly along
the intercept path 1040. A predicted rendezvous position 1050 and
direction 1060 can run parallel to the target vessel 1020 that
travels along a target path 1070 and corresponding direction
1080.
FIG. 11 shows plan view 1100 for third guidance level competency
for blocking a craft from reaching an intended position. The
platform 410 is disposed at an initial horizon boundary 1110
corresponding to interval i along an intercept path 1120 for
arrival at an intercept position 1130 in a specified travel
direction 1135. At the intercept position 1130, the platform 410
can detect and deter objects within a lethal range radius 1140. An
adjacent horizon 1150 at the next interval i+1 corresponds to a
position not directly along the intercept path 1120. The intercept
position 1130 is disposed between a defended asset 1160 and the
craft 1170 to be blocked. By traveling in a direction 1180 towards
the defended asset 1160, but reaches the radius 1140 at a range
position 1190 for deterrence by the platform 410.
FIG. 12 shows plan view 1200 for fourth guidance level competency
for search and patrol. The platform 410 is disposed at an initial
horizon boundary 1210 corresponding to interval i along an
intercept path 1220. An adjacent horizon 1230 at the next interval
i+1 corresponds to a position not directly along the intercept path
1120, which meanders around obstacles 650 and through regions 1240
for investigation.
FIG. 13 shows block diagram view of a circuit 1300 for the Route
Planning Layer 170 for Guidance. An optimizer 1310 initiates along
guide parameters path 1320, such as gain constants K.sub.1,
K.sub.2, K.sub.3, t.sub.c and .psi. to an on-board simulator 1330
and cycle along an iteration return path 1340. The circuit exhibits
directionality 1350 upon optimization to be submitted to the
navigation layer 180. The cycle 1300 optimizes parameters to
maintain a minimum closest point of approach (CPA) relative to an
obstacle in the local horizon, as well as to minimize time. The
simulator 1330 integrates over the control horizon using an assumed
guidance law with parameters, such as minimum CPA to any object
during trajectory, time-to-horizon and final states at horizon.
FIG. 14 shows plan view 1400 of a guidance layer for composite
navigation along a trajectory with an NEDL frame 940. The platform
410 is disposed at an initial bearing position 1410 and travels
along a path 1420 towards an end goal 1430, while avoiding
obstacles 650. The path 1420 incorporates composite navigation for
avoidance, homing and shaping using optimizer-selected weighting
factors. From the initial bearing position 1410 travels in a
straight initial bearing until a course correction horizon 1440
that initiates a direction change set at a selected time.
Control law solutions for terminal conditions in vector form
include all dimensions such as throttle acceleration to be used for
target intercept (i.e., offset rendezvous). The optimal control u
can be expressed as:
.function..function..function. ##EQU00001## where V is missile
speed, R is range-to-go to the next waypoint or target point,
K.sub.1 and K.sub.2 are respective guidance gains, {circumflex over
(r)} is the unit vector for the line-of-sight to the target point,
{circumflex over (v)} is the unit vector along the current velocity
and {circumflex over (v)}.sub.f is the unit vector for the desired
final velocity orientation. The commanded accelerations can be
expressed without throttle control.
Bearing guidance includes an initial bias phase, such as in a P
frame, expressed as:
.cndot. ##EQU00002## where the vector array includes the middle
term such that: a*=[k(.psi.*-.psi.)], (3) where k is a user
selected gain, .psi.* and .psi. are current bearing angles of the
vehicle, and an obstacle avoidance component, also in the P frame,
expressed as:
##EQU00003## where
##EQU00004## is a vector of acceleration commands necessary to
avoid obstacles.
An optimization or rule-based set algorithm can be used to
determine the parameters that minimize the time to the goal point,
maintains minimal separation (i.e., the CPA) around known
obstacles, and satisfies any heading constraints along the path.
Such constraints represent known heading angle constraints due to
wave motion in certain regions. One goal includes avoiding the
preplan of the route to improve safe maneuver. FIG. 13 for route
planning shows the concept for the RPL 170.
The navigation layer 180 is responsible for generating commands for
the heading, heading rate, or acceleration to the control layer
190. The navigation layer 180 uses the navigation algorithm
parameters 175 selected by the RPL 170 in computing these heading
rates or heading schedule. FIG. 14 shows the overview of the
procedure for the navigation layer 180, including an initial
bearing phase 1410 from which the vehicle 410 turns to a bearing
and holds until reaching position 1440 at a designated time t.sub.c
from the RPL 170.
Composite navigation provides for obstacle avoidance and path
shaping using the parameters from the RPL 170. Depending on the
implementation on the USV platform 410, the navigation law can be
cast as a commanded acceleration as shown:
.times..times..times..times..times..times..times..times..times..times..ti-
mes..times..times..times.
.times..times..times..times..times..times. ##EQU00005## where T and
T.sup.T are matrices and their respective transforms.
The commanded accelerations are developed in the reference frame
defined by the NEDL frame 940. The three terms provide three,
different aspects of the guidance law are blended by a set of three
gains contained in the three gain tensors: {tilde over (K)}.sub.1,
{tilde over (K)}.sub.2, {tilde over (K)}.sub.3. These gain tensors
are formed by: {tilde over (K)}.sub.i=K.sub.i[I] where [I] is the
identity tensor and K.sub.i is a blending weighting scalar for each
component of the composite law. The weightings are set by the RPL
170. For generality, the gain may be scheduled in time t such that:
K.sub.i=f(t), (5) where time-varying function f is provides rate of
gain change.
The three elements of the composite guidance include: .sub.goal
which is a goal and orientation control algorithm designed to reach
goal points on the local horizon, .sub.bearing is the acceleration
required to follow the bearing, and .sub.avoid is the acceleration
to avoid objects and environmental conditions, such as the need to
approach large waves at a specified angle. The specifics of how
these commanded accelerations are computed may be developed in many
different substations. The latter two commands are naturally
developed in the P frame which has its x axis aligned with the
current velocity vector and its z axis aligned with the local
gravity vector. T.sub.P2NEDL is the transformation from the P frame
to the NEDL frame.
Typically, the values for bearing guidance and the avoidance
guidance would be terminated at some point in the local horizon.
For instance, the bearing guidance computation can be set to zero
in response to the specified time of the RPL 170 to follow the
bearing has been exceeded. Similarly, the acceleration to avoid
objects or environmental conditions would reach a zero value once
the time-to-go to the obstacle or environmental conditions has gone
negative. This implies that the object is behind the USV platform
410 in terms of the current direction of travel.
The guidance or navigation layer 180 generates commands for the
heading, heading rate and acceleration to the control layer 190.
Guidance parameters are obtained from the latest route planning
update, and measurements from the IMU and GPS are processed using
guidance equations to compute acceleration or heading rate. The
navigation law can be expressed as a heading rate {dot over
(.psi.)} for in-flight composite guidance:
.psi..times..times..times..times..times..times..times..times..times..time-
s..times..times..times. .times..times..times..times..times.
.times..times..times..times..times..times..times..times.
##EQU00006## where V is current speed of the vessel. The first term
represents terminal conditions control; the second term determines
initial bearing phase and the third term provides obstacle
avoidance. The gain tensor is represented by the form:
##EQU00007## where n is integer 1, 2 or 3. The sign of the
commanded heading rate can be computed from the vector relationship
on the right hand side of eqn (7). This approach has the benefit of
interfacing with existing autopilots for vessel operations.
FIG. 15 shows a plan view of object cluster process 1500 using
fuzzy logic to cluster closely spaced objects. This process can be
used to augment obstacle avoidance. A first complex object 1510
includes track items A, B, C, D. A second complex object 1520
includes track items F, G, H, and J. A third complex object 1530
includes track item E. A fourth complex object 1540 travels along a
moving vector 1545. Obstacle tracks can be used as vertices for
avoidance logic.
FIG. 16 shows tabular view of Complex Object Map 1600 to Obstacle
Tracks. The first column 1610 identifies the complex object from
the cluster process 1500. The second column 1620 identifies the
corresponding object tracks. The third column 1630 lists the
corresponding vertex for avoidance. The fourth column 1640
concludes the closest point of approach (CPA) buffer for each
object.
FIG. 17 shows a plan view of goal seeking object avoidance 1700.
During operations, the divert angle to each object (port L or
starboard R) can be computed. The resulting miss metric (MM) or of
all other objects can be computed for each maneuver. A platform
1710 travels in an initial direction 1715 with a divergence
direction 1720 for avoidance acceleration. A stage goal 1730 is
disposed at the horizon edge 1560 (representing terminal
conditions) traveling in the direction 1565.
Obstacles K, L and M lie disposed between the traveling platform
1710 and the goal 1730. No avoidance is required for the K
obstacle. However, the L obstacle can be circumvented by dash-line
divert directions to port, and the M obstacle can be circumvented
by dash-line divert direction to starboard. Upon passing the M
obstacle through the divert maneuver, a goal angle 1740 directs the
platform 1710 to the goal 1730. A plot 1750 illustrates the
relation between the abscissa 1760 as the closest point of approach
and the ordinate 1770 as the miss metric. Diversion to expand
approach distance around an obstacle is maintained to equal the
miss metric reaches fifty-feet in this example.
FIG. 18 shows a tabular view of a selection matrix 1800 for
goal-seeking object avoidance. The columns include lists of
obstacle objects K, L and M, the divert direction (port and
starboard), the miss metric to the respective obstacles, the
average and the angle to target. The diversion from L to port is
selected with the corresponding angle from target circled.
FIG. 19 shows graphical view of a plot 1900 for optimizer
navigation substantiation. An exemplary docking mission involves a
jet-ski craft (starting at the origin) to a pier with a single
waypoint and an obstacle blocking a straight path. The abscissa
1910 represents downrange distance from the origin and the ordinate
1920 represents cross-range distance. An obstacle 1930 is disposed
between the origin and the pier 1940 that represents the goal. A
known-obstacle path 1960 enables avoidance of the obstacle 1930
assuming initial availability of this information to ensure an
imposed 50-foot clearance. A pop-up path 1970 provides an alternate
route limited to 1/3g (one-third gravitational acceleration)
maneuverability limit assuming a delay in obstacle detection.
FIG. 20 shows a block diagram 2000 of an adaptive
proportional-integral-differential (PID) controller for boat
lateral command. Desired dynamics 2010 are provided to a predicted
response algorithm 2015 to generate a response fed to a first sum
operator 2020 for receipt to a rule-set 2025 for adjusting gains
for receipt into a controller 2030. A signaler 2035 provides
inputs, such as yaw, yaw rate, heading angle, speed and
environmental conditions at the previous instrumentation sample to
the algorithm 2015 through a first gate 2040. A periodic impulse
rudder input 2045 is provided to the algorithm 2015 through a
second gate 2050.
The controller 2030 provides an exemplary implementation of the
control layer 190 that includes proportional (P), integral (I) and
differential (D) functions, receiving through a second sum operator
2055 error correction from feedback (negative) from the output and
commanded heading rate (positive) from the signaler 2035 as the
setpoint. Outputs from the functions are summed in a third sum
operator 2060 as boat true-dynamics output 2080, such as rudder
command in degrees. The summation output represents achieved yaw,
yaw rate and heading rate for the boat. This output 2080 is
returned as feedback through a third gate 2090 to the first sum
operator 2020, as well as to the second sum operator 2055.
FIG. 21 shows a graphical view of a response plot 2100 for an
adaptive PID controller used in boat lateral command. The abscissa
2110 represents time and the ordinate 2120 represents yaw rate. A
first damped sinusoidal curve 2140 represents an initial unadjusted
response to a kick with high peak and rapid attenuation as shown by
lower envelope decay 2145. A second damped sinusoidal curve 2150
represents an adjusted simulation response with upper envelope
decay 2155. The second curve 2150 corresponds to a time to first
zero-axis crossing 2160 occurring at about 3.6 seconds and a
desired peak response 2170. PID gains K.sub.p, K.sub.i and K.sub.d
(for proportional, integral and differential, respectively) can be
adjusted to achieve these conditions.
Various exemplary embodiments provide the USV platform 410 with the
ability to autonomously conduct a broad spectrum of missions
equivalent to those which a commander might expect of a human
operated surface vehicle. The ability to do this autonomously
provides the commander with an expanded combat and operations
capabilities without the attendant growth in manpower on the ship.
The first generation of USV requires an additional complement of
five-to-ten people to conduct operations of the platform 410.
Various exemplary embodiments decrease manning requirements and
augment combat and operational capability for the command platform
(mother ship) 330 operating a USV platform 410, increased
continuous functionality because human intervention becomes
optional such that and the systems do not degrade by fatigue from
human operation, and increased functionality because the automatic
system are able to perform missions that humans cannot
accomplish.
Various exemplary embodiments provide the use of layers of
automation with each layer dealing with the total problem at
different levels of abstraction and time horizon. This process
enables the mimicry of the human capability for strategic decision
making as compared to the tactical decision making. By parsing out
lower-level considerations for the higher strategic levels, the
amount of computational time is greatly reduced. This approach
thereby reduces complexity of the software. The minimum number of
logic paths operating at each layer reduces testing time necessary
to identify and correct the logic errors for ensuring safe and
efficient operation, as compared to branch-and-tree approaches.
Conventional alternatives of exemplary embodiments include high
personal attention for continued use of manned remote control
operations of the USV platform 410. This has the attendant problems
of increased manning requirements on the command platform 330,
lower performance caused by operator fatigue, and lower total
performance from communication lags due to fleet bandwidth
limitations and human limitations. An alternate approach to
automation is to use the branch-tree method as seen in applications
such as chess programs. This approach is very computational
intensive despite the game-playing characteristics, thus
necessitating only a very limited need for situational awareness as
compared to real world operations.
While certain features of the embodiments of the invention have
been illustrated as described herein, many modifications,
substitutions, changes and equivalents will now occur to those
skilled in the art. It is, therefore, to be understood that the
appended claims are intended to cover all such modifications and
changes as fall within the true spirit of the embodiments.
* * * * *